The method of elimination is iterative. One looks at the answers, determines that several answers are unfit, eliminates these, and repeats, until one cannot eliminate any more. This iteration is most effectively applied when there is logical structure between the answers – that is to say, when by eliminating an answer one can eliminate several others. In this case one can find the answers which one cannot eliminate by eliminating any other answers and test them alone – the others are eliminated as a logical consequence; this is the idea behind optimizations for computerized searches when the input is sorted – as, for instance, in binary search.
In order for the method to work it is necessary to list all possible, even improbable, possibilities. Any omissions render the method invalid as a logical method.
A process of elimination can be used to reach a diagnosis of exclusion. It is an underlying method in performing a differential diagnosis.